Mirage: An RNS-Based Photonic Accelerator for DNN Training

TL;DR

Mirage is a novel photonic DNN training accelerator that uses RNS to achieve high-precision, energy-efficient training, outperforming traditional systolic arrays in speed and power consumption.

Contribution

This work introduces a new RNS-based micro-architecture and dataflow for photonic tensor cores enabling high-precision DNN training with improved energy efficiency.

Findings

Achieves over 23.8x faster training than systolic arrays.

Consumes 42.8x less power with comparable accuracy.

Provides high energy efficiency without sacrificing precision.

Abstract

Photonic computing is a compelling avenue for performing highly efficient matrix multiplication, a crucial operation in Deep Neural Networks (DNNs). While this method has shown great success in DNN inference, meeting the high precision demands of DNN training proves challenging due to the precision limitations imposed by costly data converters and the analog noise inherent in photonic hardware. This paper proposes Mirage, a photonic DNN training accelerator that overcomes the precision challenges in photonic hardware using the Residue Number System (RNS). RNS is a numeral system based on modular arithmetic, allowing us to perform high-precision operations via multiple low-precision modular operations. In this work, we present a novel micro-architecture and dataflow for an RNS-based photonic tensor core performing modular arithmetic in the analog domain. By combining RNS and photonics, Mirage provides high energy efficiency without compromising precision and can successfully train state-of-the-art DNNs achieving accuracy comparable to FP32 training. Our study shows that on average across several DNNs when compared to systolic arrays, Mirage achieves more than $23.8\times$ faster training and $32.1\times$ lower EDP in an iso-energy scenario and consumes $42.8\times$ lower power with comparable or better EDP in an iso-area scenario.